Because I use them for testing, so I generally want to uniformly generate samples within a particular size of tree.
(generating random trees the naive way tends to produce the occasional huge tree amidst huge numbers of tiny trees)
(generating random trees the naive way tends to produce the occasional huge tree amidst huge numbers of tiny trees)
Concrete examples: Docker containers, PDF parts, POSIX filesystems, Scheme sources, SVGs, x86 objects, etc.